基于Rough熵的决策表约简算法研究

The Algorithm to Attribute Reduction Besed on Rough Entropy in Decision Table

知识约简是Rough集理论研究中的核心内容之一,已证明求决策表的最小约简是NP-hard问题,解决办法一般是采用启发式算法.本文分析了信息熵在Rough集理论中的应用,利用Rough熵定义了决策表中条件属性的相对重要度,讨论了其性质,在此基础上提出了一种用于计算决策表中的相对约简的启发式算法.

The attribute reduction is a core problem of the rough set theory,It has been proven that computing the optimal reduction of decision table is a NP-hard problem.In the paper here,the application of rough entropy in rough sets theory is analysed,the uncertainty measure of the importance of attribute in decision table is proposed;then,a heuristic algorithm based on rough entropy for reduction of knowledge is proposed.